Search Service Interoperability

[Eliot Christian <echristi@usgs.gov>]

On November 17, 2003, the OASIS e-Government Technical Committee
publicly solicited comments on the initial version of the document
concerning "Search Service Interoperability". Comments were sent
to me and I have compiled a summary showing 50 comments received
from about 15 people. That summary and my draft responses are at
http://www.oasis-open.org/apps/org/workgroup/egov/download.php/4651/wd-egov-searchservice-01-comments.html

Many people took the time to express support and to offer words
of encouragement. These, of course, are very much appreciated.

As an overall assessment of the comments received, I agree with
Mike Taylor who wrote on 22 December 2003:

   [...] Lots of the comments ask that the document be expanded
   to cover more ground, make greater demands, encompass more
   areas--in other words, some commenters want this to be a
   different document from the one it is.  My personal feeling
   that you should rather look on these requests as opportunities
   to find the directions for follow-up or otherwise related
   documents.  I'd like to see the one you've already written
   stay as small and focussed as it currently is.  The very
   fact that it's attracted so many comments shows that, in
   its current form it's not too intimidating for people to
   take the time to read.  To my mind, that's Desideratum Number
   One. Make it bigger and you'll lose half your audience.

   (To be clear -- that's not _at all_ to say that the other
   areas raised are not important.  Just that we're more likely
   to make progress by taking a sequence of small steps than a
   single huge leap.)

To me, the most surprising comments were those that challenged
the desirability of interoperability itself. These came from
people apparently involved in records management and archiving.
For example:

   [...] NARA has concluded that the interface/display is so
   critical to understanding archival materials in their
   hierarchical context that Z39.50 was not desirable. [...]
   The overwhelming majority of records are not produced for
   publication, but to support a specific transaction or series
   of actions, and they are seldom communicated beyond a
   relatively narrow circle of uses related to the original
   purpose including management, audit, etc.

Although I believe that point is debatable, interoperability
within the records management community is not really a matter
we can debate within the OASIS e-Government TC. My proposed
action is simply to delete from the document the few references
to archives and records management.

Various commenters focused on approaches available for creating
collections of information, such as "harvesting", and on the
technologies for representing semantic relationships, such as
RDF. These are certainly interesting but quite separate from
the recommendation about search interoperability through a
common search service. A new section, 3.6., was added to note
these approaches as compatible with the recommendation although
not quite in scope.

Text concerning access control in ISO 23950 was added into
section 2 of the document.

Text concerning extension mechanisms in ISO 23950 was added
into section 3.2.

Many other comments raised points that are discussed in my
proposed responses but annotated as: "This clarification is
background information. No action taken in the document."

I look forward to further discussions and on taking this
document to its next step.

Eliot